net/mlx5: use do_aux_work for PHC overflow checks

The overflow_work is using system wq to do overflow checks and updates
for PHC device timecounter, which might be overhelmed by other tasks.
But there is dedicated kthread in PTP subsystem designed for such
things. This patch changes the work queue to proper align with PTP
subsystem and to avoid overloading system work queue.
The adjfine() function acts the same way as overflow check worker,
we can postpone ptp aux worker till the next overflow period after
adjfine() was called.

Reviewed-by: Dragos Tatulea <dtatulea@nvidia.com>
Signed-off-by: Vadim Fedorenko <vadfed@meta.com>
Acked-by: Tariq Toukan <tariqt@nvidia.com>
Link: https://patch.msgid.link/20250107104812.380225-1-vadfed@meta.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c b/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c
index 4822d01123b45d575eca40f2746b4e54408bc680..d61a1a9297c904f1184c601ea676f64617791b68 100644 (file)
--- a/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c
@@ -322,17 +322,16 @@ static void mlx5_pps_out(struct work_struct *work)
        }
 }
 
-static void mlx5_timestamp_overflow(struct work_struct *work)
+static long mlx5_timestamp_overflow(struct ptp_clock_info *ptp_info)
 {
-       struct delayed_work *dwork = to_delayed_work(work);
        struct mlx5_core_dev *mdev;
        struct mlx5_timer *timer;
        struct mlx5_clock *clock;
        unsigned long flags;
 
-       timer = container_of(dwork, struct mlx5_timer, overflow_work);
-       clock = container_of(timer, struct mlx5_clock, timer);
+       clock = container_of(ptp_info, struct mlx5_clock, ptp_info);
        mdev = container_of(clock, struct mlx5_core_dev, clock);
+       timer = &clock->timer;
 
        if (mdev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR)
                goto out;
@@ -343,7 +342,7 @@ static void mlx5_timestamp_overflow(struct work_struct *work)
        write_sequnlock_irqrestore(&clock->lock, flags);
 
 out:
-       schedule_delayed_work(&timer->overflow_work, timer->overflow_period);
+       return timer->overflow_period;
 }
 
 static int mlx5_ptp_settime_real_time(struct mlx5_core_dev *mdev,
@@ -517,6 +516,7 @@ static int mlx5_ptp_adjfine(struct ptp_clock_info *ptp, long scaled_ppm)
        timer->cycles.mult = mult;
        mlx5_update_clock_info_page(mdev);
        write_sequnlock_irqrestore(&clock->lock, flags);
+       ptp_schedule_worker(clock->ptp, timer->overflow_period);
 
        return 0;
 }
@@ -852,6 +852,7 @@ static const struct ptp_clock_info mlx5_ptp_clock_info = {
        .settime64      = mlx5_ptp_settime,
        .enable         = NULL,
        .verify         = NULL,
+       .do_aux_work    = mlx5_timestamp_overflow,
 };
 
 static int mlx5_query_mtpps_pin_mode(struct mlx5_core_dev *mdev, u8 pin,
@@ -1052,12 +1053,11 @@ static void mlx5_init_overflow_period(struct mlx5_clock *clock)
        do_div(ns, NSEC_PER_SEC / HZ);
        timer->overflow_period = ns;
 
-       INIT_DELAYED_WORK(&timer->overflow_work, mlx5_timestamp_overflow);
-       if (timer->overflow_period)
-               schedule_delayed_work(&timer->overflow_work, 0);
-       else
+       if (!timer->overflow_period) {
+               timer->overflow_period = HZ;
                mlx5_core_warn(mdev,
-                              "invalid overflow period, overflow_work is not scheduled\n");
+                              "invalid overflow period, overflow_work is scheduled once per second\n");
+       }
 
        if (clock_info)
                clock_info->overflow_period = timer->overflow_period;
@@ -1172,6 +1172,9 @@ void mlx5_init_clock(struct mlx5_core_dev *mdev)
 
        MLX5_NB_INIT(&clock->pps_nb, mlx5_pps_event, PPS_EVENT);
        mlx5_eq_notifier_register(mdev, &clock->pps_nb);
+
+       if (clock->ptp)
+               ptp_schedule_worker(clock->ptp, 0);
 }
 
 void mlx5_cleanup_clock(struct mlx5_core_dev *mdev)
@@ -1188,7 +1191,6 @@ void mlx5_cleanup_clock(struct mlx5_core_dev *mdev)
        }
 
        cancel_work_sync(&clock->pps_info.out_work);
-       cancel_delayed_work_sync(&clock->timer.overflow_work);
 
        if (mdev->clock_info) {
                free_page((unsigned long)mdev->clock_info);

diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h
index ea48eb879a0f5265128bfcf7219ff0ac0cf4c61b..fed666c5bd1632f73e4ca72cbe019cfd8a283f1b 100644 (file)
--- a/include/linux/mlx5/driver.h
+++ b/include/linux/mlx5/driver.h
@@ -691,7 +691,6 @@ struct mlx5_timer {
        struct timecounter         tc;
        u32                        nominal_c_mult;
        unsigned long              overflow_period;
-       struct delayed_work        overflow_work;
 };
 
 struct mlx5_clock {